This week’s AI developments show organizations moving further into the practical work of implementation. Companies are redesigning processes around AI, building models for specific business functions, and giving agents greater responsibility inside enterprise systems. Those developments are also creating new questions about governance, workforce planning, healthcare oversight, and how much autonomy organizations should give AI. For business leaders, the focus is increasingly on building operating practices that allow AI to produce measurable results without introducing unnecessary risk.
1. Stop Automating Old Processes. Redesign Them Around AI.
A new Harvard Business Review article argues that many companies are limiting their AI returns by applying the technology to individual tasks inside existing processes. Instead, organizations should reconsider the entire workflow and determine how work should be structured when AI is available from the beginning. Research cited by HBR indicates that while generative AI adoption is widespread, enterprise-wide financial results remain difficult to achieve.
The idea has applications across many functions. A finance team, for example, could rethink the complete process for reviewing transactions, investigating exceptions, preparing analysis, and securing approvals rather than simply using AI to draft a report at the end. Customer service teams could similarly redesign how requests are categorized, researched, escalated, and resolved.
Read the full story on Harvard Business Review
2. Salesforce and Nvidia Build an AI Model Specifically for CRM
Salesforce and Nvidia introduced Koa, an open reasoning model trained specifically for customer relationship management. Rather than operating as another general-purpose model, Koa is designed for tasks involving sales, marketing, and customer service. The development reflects growing interest in AI models designed around particular business functions and enterprise data.
A CRM-focused model could help sales teams research accounts, identify potential opportunities, prepare customer communications, or determine appropriate next steps. Customer service organizations could use similar capabilities to interpret requests, retrieve relevant information, and help agents resolve cases more efficiently.
Read the full story on TechCrunch
3. Companies Are Building New Guardrails for AI Agents
As AI agents gain the ability to take actions across business systems, companies are creating additional layers of oversight. Fortune reports that Cisco, Intuit, Workday, and ServiceNow are among the organizations working on ways to monitor agents, restrict their permissions, and intervene when systems behave unexpectedly.
The issue becomes particularly important when agents move beyond research and content generation. An agent that can modify records, communicate with customers, access financial information, or trigger business processes requires different controls than a chatbot that only answers questions. Companies may need spending limits, permission boundaries, activity logs, approval requirements, and automatic shutdown mechanisms.
Read the full story on Fortune
4. Businesses Need to Prepare for Several Possible AI Workforce Outcomes
A new report from The Conference Board outlines four ways AI could reshape the U.S. workforce, ranging from gradual employee augmentation to substantial displacement. The organization projects that within three years, 60% to 70% of jobs in the cognitive workforce could involve collaboration between humans and AI, compared with 15% to 25% involving human-only work.
For employers, the uncertainty makes workforce planning especially important. Companies can identify which roles are likely to use AI extensively, determine which skills employees will need, and establish training programs before job requirements change significantly. AI literacy may increasingly become part of professional development across finance, marketing, operations, HR, IT, and other knowledge-based functions.
Read the full story on The Conference Board
5. AI Governance Policies Are Not Keeping Up With Actual Deployment
New EY research highlights a gap between formal AI policies and what happens when organizations deploy AI in practice. Although 98% of surveyed senior AI executives said their organizations have formal AI governance policies, 47% acknowledged that their organizations have bypassed those processes for urgent deployments. Among organizations using agentic AI, 26% said they cannot detect unauthorized agents operating internally.
The consequences are becoming tangible. More than one-third of respondents said their organizations had experienced an AI incident or failure that produced a materially negative impact, including operational disruption, financial damage, data loss, or harm to the organization’s reputation. For AI leaders, the findings suggest that written policies need to be accompanied by technical controls, agent inventories, monitoring, and clear accountability.
6. 2026 Corporate AI Talent Study – Infographic
The first annual Corporate AI Talent Study examines how organizations are preparing employees for the AI era including talent strategy, hiring, training, skills development, and workforce impact. Based on a national survey conducted between June and August 2026 with more than 300 executives across North America, the study provides a timely benchmark for how organizations are building AI capabilities within their workforce.
Why It Matters
- Companies may realize greater AI value by redesigning complete workflows rather than automating isolated tasks.
- Purpose-built AI models are creating more specialized applications for sales, marketing, customer service, and other business functions.
- AI agents require stronger controls as they gain permission to take actions inside enterprise systems.
- Workforce planning should account for several possible AI scenarios, with human-AI collaboration likely to become a larger part of knowledge work.
- Formal governance policies are insufficient if organizations cannot identify unauthorized agents, enforce controls, and monitor AI activity in practice.
Stay Ahead with the AI Insiders Newsletter
Join more than 55,000 AI executives and innovators for exclusive insights, strategies, and trends shaping the future of AI leadership. Subscribe to the AI Insiders Newsletter.
